Ahamad Ghanbari, Baratali Fakheri and Abolfazl Tavassoli
Abstract
This research was carried out in order to simulation of wheat yield in water and nitrogen limit conditions in Shirvan region, Iran, using models of CERES-Wheat&AquaCrop. The goal of doing this research is evaluating these two models for simulating wheat yield and biomass. In order to obtain necessary data for calibrating and validating model, a field experiment was conducted as split plot on randomized complete block design with three replications in 2009-10 and 2010-11 on winter wheat (Sayonz cultivar). Irrigation treatments were composed of I1: optimum irrigation (field capacity (FC)), I2: 80% FC, I3: 60% FC and I4: without irrigation (dry farming) as the main plots. Nitrogen fertilizer was composed of N1: 0, N2: 100, N3: 175 and N4: 250 kg/ha as the sub plots. The results obtained from simulation by CERES-Wheat model showed that biomass RMSEn for years of 2009-10 and 2010-11 was equal to 5.03% and 5.30% respectively; and R2 for both years obtained 0/9920 and 0.9911 respectively. Seed yield RMSEn for years of 2009-10 and 2010-11 was equal to 8.95% and 9.11% respectively; and R2 for both years obtained 0.9817 and 0.9779 respectively. These results show that biomass and seed yield was simulated well. The results of simulation by AquaCrop model showed that model simulates biomass and seed yield optimally. So that biomass RMSEn for years of 2009-10 and 2010-11 was equal to 10.81% and 11.10% respectively; and R2 for both years obtained 0.9689and 0.9638 respectively. Seed yield RMSEn for years of 2009-10 and 2010-11 was equal to 16.75% and 14.75% respectively; and R2 for both years obtained 0.9413 and 0.9558 respectively. With due attention to results achieved, comparison and evaluation of two models on the basis of R2 and RMSEn indicators showed that CERES-Wheat model has simulated wheat biomass and yield (Sayonz cultivar) well and simulation accuracy of CERES-Wheat model is better than AquaCrop model in Shirvan region.